Thursday, January 16, 2020

The utility of directly regulating Floor Area Ratio

In my last post on the obesity of midrise apartments, I mentioned that one possible solution to what I perceive as a problem is to restrict FAR (Floor-Area Ratio) tighter and loosen setback and height limits so that buildings' density be more limited by FAR regulations than geometric regulations. I got a bit of pushback on Twitter for that, and it's quite understandable, as it seems unintuitive and even counter-productive.

Unintuitive because whatever we might think of them, setback regulations and height limits can be justified, convincingly or not, by the idea of externalities. Buildings that are too high or too close to property limits are viewed as intruding visually (if nothing else) on neighboring properties, arguably justifying regulations to reduce these impacts on other properties or on the street. FAR regulations don't really have this argument though, not in a direct way anyway.

Counter-productive because, well, FAR is the best metric for built density. So by arguing for stricter FAR regulation, that means a restriction of density... when I've been arguing (and still do) for more density usually. Isn't that the opposite of what we should want?

Well, give me some time and I'll explain why I'm arguing for this.

The role of FAR in determining land value

One thing that's important to understand is that URBAN LAND HAS LITTLE TO NO VALUE IN ITSELF.

But urban lots are very valuable, so how can I say that?

Simple. Urban land has no value, it is the building allowance tied to that lot that has value. Developers buy land in order to build valuable buildings on them and reselling them. Speculators buy land they think is underpriced or will increase in value to sell to a developer down the line. So urban land's value is ultimately determined by what developers are willing to pay for it, and what is a developer willing to pay for a lot?

There's essentially three variables that help determine what a developer will be ready to pay for an urban lot.

  1. The market value of a square meter of floor area in the neighborhood (P)
  2. The construction cost of building that square meter of floor area (C)
  3. The amount of square meters of floor area that the developer can build on that lot (A)
 So you can sum up the value of an urban lot, regardless of its actual size, by the formula:

LOT VALUE (P-C)A

Okay, profit margin should be included there as well, but for simplicity's sake, let's set it aside since it would only complicate explanations and anyway, it's pretty much a constant.

You'll notice I put "less than or equal to" in there rather than an equal sign. That's because lot value can be less than the formula, if speculators face a lot of competition and pressure to sell, they can choose to sell for less. But a developer cannot pay a lot more than he can hope to get as income from it, he would go in the red if he did that, whereas the speculator is, usually, just cashing in his profits. Sacrificing a bit of your profits is easier to do than committing to your own personal bankruptcy...

Looking at a micro view, meaning from the point of view of one economic actor (the developer), two things are essentially independent variables, the market value of housing per square meter and the construction cost of housing, which are determined by market conditions in that neighborhood and the construction industry. The one variable that can change most is the amount of square meters of floor area they can build, because it is an outcome of municipal regulation and the developer can even ask for zoning variances or changes to increase it (thus increasing the value of the lot).

The area that can be built ties directly to FAR. So, if my understanding is correct, FAR is directly proportional to the value of urban lots.

 LAND VALUE ($/M2) = LOT VALUE/LOT AREA (P-C)A/LOT AREA= (P-C) times FAR

So, the conclusions of this reasoning are the following:

  1. Lot value is directly proportional to the FAR developers expect to be able to build on it
  2. When FAR is not regulated, or when cities are open to spot zoning or FAR-raising variances, then this creates uncertainty about the ultimate FAR allowance of a lot, and thus, its value
  3. Speculators will tend to estimate optimistically the FAR that will be ultimately allowed on their lot (because they want to maximize profit), developers will tend to estimate pessimistically the FAR (because they want to minimize risk)
  4. It is crucial for reducing uncertainty and allowing speculators and developers to come to terms more readily to stabilize FAR with rigorous regulations, this helps stabilize land prices and helps the liquidity of lots
Therefore, FAR regulation, though it doesn't reduce externalities, does reduce economic uncertainty and helps facilitate deals between speculators and developers to allow the development of land.
I'll point out that FAR limits are a major part of Japanese zoning, the description of which I'm most well known. And if you agree with my assessment that the Japanese zoning system works uncommonly well for a zoning system, I don't think you can discount the role that FAR regulation plays in it.

Furthermore, it's important to understand that FAR is essentially always regulated by urban regulation and zoning.

The de facto FAR regulation of geometric regulations

Let's take a lot, 60 meters wide and 35 meters deep:

Let's suppose there are front and back setback regulations of 6 meters, and let's draw them on the lot:

Let's also suppose that there are side margin requirements of at least 3 meters on each side:

And then let's suppose there's a 6-story height limit:


The result of all these regulations is a box within which it is allowed to build, and outside of which it is illegal to build.

De facto, these geometric rules represent both:
  • A lot coverage maximum (how much of the lot can be built over)
  • A FAR maximum (representing lot coverage times the number of stories that can be built)
So there's no reason why you need an actual FAR maximum then, because it's already provided for by other rules, correct?

No, not correct.

The problem


The issue is that the lot's value is proportional to FAR, and if your allowed FAR is determined by geometric regulations, then developers willing to build up to the limit in every way will easily outbid anyone trying any other design. So basically, without an actual FAR restriction that is lower than the de facto FAR restriction of geometric rules, by setting those geometric limits, you already imposed on a developer the envelop of the building he will build. Notice how the last image's red block is nearly identical to those obese midrises I described? That is not a coincidence.

Let's continue on the previous example, the de facto regulations of the building are the following:
  • Maximum lot coverage: 54 meters wide by 23 meters deep, 1 242 square meters on a 2 100 square-meter lot, so 59% (let's round that up to 60%).
  • Maximum FAR: 60% times 6 stories equals 360% FAR
What happens if a developer who read my previous blog decided to make a 16-meter deep building only to maximize exterior walls and provide for more bedrooms per floor area? Well, his maximum FAR would be just about 40% by story, so 240% total.

That means that the FAR optimizing developer building obese midrises is going to massively outbid the slim building developer, his bid would be 50% higher because lot value is proportional to FAR. 

Alternatively, that would mean the thinner building would need to be able to sustain a much higher price per square meter to be viable. How much higher? That depends on how expensive the land is, so I'll answer by a table:


So, as we can see from this table, the more expensive land is, the more the premium that will have to be paid to afford the thinner midrise and its units with more bedrooms. Which means in suburban areas, building less than the maximum FAR is actually doable, it will increase prices by only 10% or so by square meter... but for areas with high land costs, the difference becomes incredibly high, making the thinner option completely non-viable.

So, to sum up, by restricting FAR through geometric regulations rather than directly, you:
  1. Essentially impose thick, deep buildings that fill up this entire space in urban areas.
  2. Where land prices are high, you get almost exclusively 1-bedroom apartments as a result (except where lots are smaller and have detached buildings only), anything else becomes much too expensive to build and buy because it would require to cut down the FAR and thus have much higher prices per square meter
  3. In suburban areas, this problem is much reduced, and so you'll get more flexibility on building shape and size.
Note that other regulations can also affect what is built. For example, minimum parking requirements famously restrict the amount of units that can be built by the amount of parking that can be provided for them on the lot. So in areas with little FAR restriction and high parking requirements, you're likely to get extremely big units, as it's the most profitable form of building that can be built... which likely explains the insane housing consumption of most of America.

A proposal

So my proposal after this reflection is the following:
  1. FAR needs to be regulated directly, and it needs to be done in a strict manner, not easily modifiable by zoning variances or spot zoning, in order to create certainty with regards to land valuation
  2. Maximum FAR ought to be at most 70 or 80% of "de facto maximum FAR", which is the FAR calculated as the maximum allowed if all geometric dimensions were pushed to the limits allowed by geometric regulations (height limits, setbacks, margins), so that developers are more free to adopt building shapes adapted to people's needs rather than just build thick boxes.
  3. Geometric regulations can be kept to consider externalities, and should still be modifiable with variance demands, flexibility on geometry, inflexibility on FAR
  4. Overall, allowed FAR should be increased by right in all neighborhoods to flood the market with available, untapped FAR, which should reduce the land value costs per square meter of floor area.
  5. In this regulation, there is one exception, there should be a low-rise zone with a 4-story limit that should be strictly respected and FAR designed accordingly, because low-rises are much, much cheaper to build than midrises and high-rises. Once a zone is upzoned from low-rise to mid-rise, the jump in FAR allowed should be significant to compensate that increase in construction costs.

Monday, January 13, 2020

The midrise obesity crisis in North America

Well, it's been a long while since my last post. A combination of a busy life, a feeling that I had said most of what I felt comfortable saying and a desire not to repeat myself contributed to this silence. That being said, I have just found something that inspired me to break my silence.

Note, in this blog, I will use metric, here are easy, rough conversion factors:

1 square meter = 11 square feet
3 meters = 10 feet

The modern mid-rise mini-boom

In most North American cities, a new type of building has started to emerge, generally designed to offer urban condos for young professionals in search of more urban housing accommodations. Due to the homogenization of the market, these often share architectural traits and similar look, leading to some protesting about their "cookie cutter" nature, being built in all cities without respecting the local vernacular architecture.

Here is just one article criticizing the style: Why America’s New Apartment Buildings All Look the Same

If you live in a North American city, these will look familiar to you... no matter in what city you live
Now, I'm not going to criticize them too much on the look, because:
  1. Every new architectural style created a backlash when it was introduced, before later on being cherished by the same who criticized them... see for example "browstone" buildings in New York, first seen as soulless atrocities, now absolutely beloved by architects and urbanists.
  2. These represent a new boom in urban housing that had been underdeveloped for decades, so I'm of the advice we shouldn't start objecting to something necessary just because it's not to the aesthetic taste of all.
That's not to say that this style of building is perfect, far from it. And this is why I'm writing this blog post, to criticize them for something I've yet to see addressed by most people, a problem of that style of development this is way more important than just aesthetic considerations and that I believe has major repercussions on the kind of housing provided by these new developments.

Without further ado... These new apartment buildings are just too fat!

Now that I have your attention, let me provide a few examples of apartment buildings in many different countries to illustrate the point.

Fit, slim modern and traditional apartment buildings around the world: providing family housing at urban densities

Haussmanian apartment buildings in Paris, the depth of the building measured here is 9,7 meters
Vienna traditional Euroblocs, depth of 12 meters

These more recent Euroblocs in Prague have a depth of 16 meters
Warsaw older apartment buildings, depth of 10 meters
Singapore, urban apartment buildings, depth of 12 meters
Project from Urban Renaissance in Tokyo, depth of 9 meters
Manshon projects in Sapporo's suburbs, depth of a bit less than 15 meters
Apartment buildings in Uppsala, a midsize Swedish city, depth of 14 meters
Boston's Methunion Manor coop buildings, depth of 11 meters
Note that the depth of these buildings from Europe, Asia and North America, meant to provide accommodations to all kinds of households in urban (or even suburban) areas varies from 8 to 15 meters, with most being from 9 to 12 meters.

Now, let's look at some recent midrise condos in North American cities, shall we?

Fat, obese new urban apartment buildings of North America


New condo buildings in the Griffintown neighborhood of Montréal, depth of 21 meters
Chicago, depth of 19 meters
Seattle, depth of 20 meters
Dallas, depth of 21 meters
Los Angeles, depth of nearly 22 meters
There are just an handful of examples, but with your knowledge of your own city and Google Maps, I invite you to locate new developments and also measure the depth of these new developments. They tend to average around 20-22 meters.

Why it matters

So, now that I've hopefully provided enough proof of a difference between the "slim" apartment buildings of Europe, Asia and old North American developments and the "thick" apartment buildings of the current midrise urban housing boom, one might say: 

"Well, ok, but what does it matter? Doesn't it just mean greater lot coverage and density for a given number of story?"

Well, yes it does mean greater lot coverage, but it does matter a lot if one's objective is to make sure all households can find accommodation in urban areas. That's because of windows. Either because of regulation or of market demand, it is generally accepted that two types of rooms in an unit require windows on the outside: bedrooms and the living room. The larger an household is, the more bedrooms they need, at least, according to modern living standards. So, families obviously require apartments with more bedrooms, one for the parents to share, and at least one bedroom for every two children (and ideally, if children have to share one bedroom, it should be bigger than if they each had their own bedroom).

So, the more bedrooms an unit has, the more exterior wall it must "consume". The floor area doesn't increase proportionally however, if a 1-bedroom unit for 2 people is about 60 square meters, a 2-bedroom unit for a family of 2 parents and 1 child with 70 or 75 square meters is sufficient, an increase of only about 20% to deal with a family that is 50% bigger. Likewise, a 3-bedroom unit can make do with 85-90 square meters, and accommodate a family with 2 or even 3 children.

To sum up easily, with basic North American urban standards:


Visually, let's look at the floor area consumption and the exterior wall consumption of these family types:
We can notice that the exterior wall consumed increases much faster than the floor area as families grow more numerous. This suggests a new metric that ought to be considered, the Floor-Area-to-Exterior-Wall ratio.

Let me calculate it for each family's preferred housing unit and for three types of apartment buildings, the slim traditional apartment building with a depth of 10 meters, the thick traditional apartment building with a depth of 15 meters and the modern obese apartment building with a depth of 20,5 meters. For simplicity's sake, I'll presume an infinitely long apartment, it won't change results much.

 *****
ADDED ON 2020-01-14

Alon Levy on Twitter remarked that living space by person in the US tends to be significantly higher than the numbers I selected. It's true that I used minimum floor area sizes reflected in Montréal and a lot of cities with expensive real estate... what happens if I adjust the numbers a bit for higher to take that into account?



So even adjusting for bigger apartments, the conclusion remains the same. Now, luxury apartments or suburban apartments obviously go beyond that minimum, but I don't think these are relevant or typical of what could be build but cannot when buildings start getting too thick.
 
END OF ADDENDUM
***** 

What this metric reveals is that the modern obese apartment building doesn't provide near enough exterior wall to provide the desired housing units of households with more than 2 people in them. As a result, if someone wants a 2-bedroom or a 3-bedroom unit, they are going to consume a lot more floor area than they would actually need. This has the consequence of increasing the cost of such units needlessly and by a lot.

For example, imagine you have a modern apartment building with a floor area to exterior wall ratio of 10,5, and you want to make sure all units in it have 2 bedrooms, then while these units would only need 72 square meters of floor area in a properly shaped building, you would be forced to provide each unit with around 95 square meters of floor area. That's nearly 30% more... and that also means that the cost will be 30% higher. So if you'd be able to make and sell 72-square meter units for 300 000$, the 95-square meter unit would need to sell for 400 000$ instead. That is prohibitively more expensive for a lot of people.

I could make geometrical explanations, but I think this is telling enough.

Consequences of obese apartment buildings

The result of this shape is that the vast majority of units in these new buildings are lofts or 1-bedroom units, and there are only rare 2-bedroom units or more, and they tend to be significantly more expensive than they ought to be. This makes families unable to profit from the recent boom in urban housing.

One could say that at least it responds a need, that's true, but this shape is pernicious because it can't be easily remedied. You can't easily reconfigure the interior of the apartment to accommodate more than 1-bedroom units. So you bake in a bias for these units for singles whenever you build these buildings. It's fine if it would be the occasional building, but it's too common for that. If we want to allow families to live in urban areas, we need to find a way to correct this.

Maybe we ought to implement regulation imposing a maximum ratio of floor area to exterior wall in new buildings.

Why it came about

My guess as to how it came about is that it's a result of zoning regulations and financial incentives. Most North American zoning regulations control the following:
  1. Front setback
  2. Back setback (the building can't extend to a certain distance of the property line in the back)
  3. Height limit
Meanwhile, most North American zoning regulations don't control FAR (floor area ratio), or when it does, it's not as strictly controlled as height and setbacks.

So, if you're a developer who wants to maximize profit (meaning, produce the highest floor area possible from a lot) and you have to respect strict setback rules and height limits,  then the easy solution is to build a building that occupies all the lot you are legally allowed to build on and to build the maximum number of stories.

One problem is that once developers start doing that, then all lots are priced according to that approach that maximized FAR, and so developers who would use alternative shapes that reduce FAR but increase the possibility of family-sized units are punished financially by paying for a lot priced for a higher FAR than they can build. That cuts into their profits.

A solution would be stricter FAR control in regulation (to control lot prices) and looser height and setback regulations.

Approaches to maximize lot coverage and still offer a reasonable amount of exterior walls

These last examples are buildings that also seek to maximize lot coverage while still taking into account the need for exterior walls to provide family-sized housing units, through the use of peculiar forms more complex than a simple rectangle.
Rare example in North America, a condo building from 2013 in Montréal. The roughly 6x6 meter cuts in the façade enable this building to have 3-BR condos that are less than 100 square meter big and selling for 280 000$ (OK, not in a particularly desired location)

Apartment buildings in Hong Kong, shaped like intersecting Hs, creating exterior walls between the arms of the building
New apartment buildings in Singapore, also using irregular shapes that look like the letter H, again, to maximize exterior walls and allow the production of more multi-bedroom units
This extremely deep apartment building in Madrid, Spain, has inner courtyards that reduce the effective depth of the building from 25 meters to just 12 meters
Some tests I made of different apartment building shapes and FAR per story of each, showing high lot coverage can be achieved while avoiding the obese apartment buildings we're building currently